Re: NetBSD 7.0 i386 mkdep fails

[David Lord]

On 6 Jun 2016 at 7:36, Mayuresh wrote:

> On Sun, Jun 05, 2016 at 07:55:21PM -0000, David Lord wrote:
> > For many years I've been using pkgsrc/sysutils/sysbuild to 
> > generate kernels, sets, release img and iso. It downloads source
> > and builds chroot so your host system configuration doesn't cause
> > complications. You can cross build or build both i386 and amd64
> > from same conf file which has separate sections for each arch.
> 
> Sounds very handy. Thanks for sharing the tip. Does it have an option to
> build just kernel?
> 
> Mayuresh.


>From my conf file I have:

#
MKX11="yes"
MACHINES="i386"
BUILD_TARGETS="tools kernel=NTP-i386_0.80 release \
iso-image install-image"
INCREMENTAL_BUILDS="NO"
CVSTAG="netbsd-7"
#

I'm not sure if tools is required as a target if only 
kernels are required.
 

David

Re: NetBSD 7.0 i386 mkdep fails

[David Lord]

On 5 Jun 2016 at 13:26, Mayuresh wrote:

> On Sun, Jun 05, 2016 at 01:17:01PM +0530, Mayuresh wrote:
> > On Sun, Jun 05, 2016 at 07:42:22AM +, co...@sdf.org wrote:
> > > > # tar tvfz syssrc.tgz | grep "usr.bin/make"
> > > > #
> > > 
> > > The easiest way to get it to work is to grab the full source sets.
> > > 
> > > I'm not sure build.sh should be checking for it, it is just testing if
> > > it's at /usr/src in a way that is unlikely to change any time soon.
> > > 
> > > I strongly suspect that after doing this, even `make depend` will Just
> > > Work.
> > 
> > It probably might solve the problem. But is it good for kernel build to
> > depend on base/other sources in any way?
> 
> While my problem is solved thanks to Robert's hint (my mistake in mk.conf
> when switching from amd64 to i386) above question still intrigues me. Will
> be glad if this could be discussed further.
> 
> Mayuresh.
> 


Hi

For many years I've been using pkgsrc/sysutils/sysbuild to 
generate kernels, sets, release img and iso. It downloads source
and builds chroot so your host system configuration doesn't cause
complications. You can cross build or build both i386 and amd64
from same conf file which has separate sections for each arch.

Slight disadvantage might be disk space usage.
$ sudo du -sm /home/sysbuild/
16569  /home/sysbuild/



David

Re: building current

[David Lord]

On 7 Feb 2016 at 3:04, Darren wrote:

> 
> 
> 
>   From: "co...@sdf.org" 
>  To: Darren  
>  Sent: Saturday, February 6, 2016 7:40 PM
>  Subject: Re: building current
>   
> On Sat, Feb 06, 2016 at 09:02:07PM +, Darren wrote:
> > I've made many attempts to compile current using 
> > https://wiki.netbsd.org/tutorials/how_to_build_netbsd-current/ for 
> > reference.  
> > I think maybe it's outdated or current has been broken for a week?  I've 
> > cvsupdated serveral times. 
> 
> What command are you running and where is it failing? I built it a few hours 
> ago.
> 
> Just the commands on that site for amd64.  I think it's something wrong with 
> the instructions on build or cvs.  I'm running sysbuild from pkgsrc as 
> recommended on another site and it's working fine.  
> I'm more used to freebsd's method of buildworld, buildkernel, installkernel, 
> installworld.  The script build.sh is new to me.
 
I used sysbuild for Netbsd-6 current on both i386 and
amd64

Now I build NetBSD-7. Twice a week for i386 and once a 
week for amd64. 

On my systems sysbuild4cron from installed sysbuild has
an error $SYSBUILD_BINDIR="/usr/pkg" instead of 
/usr/pkg/bin so I moved sysbuild to /usr/local.

When I have build failures it's usually one of:
my error in a kernel.conf, interrupted internet
connection, corrupted downloaded source file.


David
 
   

Re: NPF on raspberry pi 2

[David Lord]

On 20 Dec 2015 at 8:13, Yves Bovard wrote:

> It isn't a stupid question! The device doesn't exist. How can I create it,
> or who creates it? I first thought that "npfctl reload" or "npfctl start"
> will create this device but both complain that /dev/npf doesn't exist
> 

Hi

My kernel configuration for i386 has a line:
#pseudo-device   npf

To use npf that would need to be:
pseudo-device   npf

and then build and npf enabled kernel.

After the system has been booted from the new kernel the
next step is:
cd /dev 
MAKEDEV 
or
MAKEDEV all



David

Re: wsrc group (Xpost, was: Question: wip, git, ssh and permissions)

[David Lord]

On 26 Aug 2015 at 13:32, Ottavio Caruso wrote:

 On 25 August 2015 at 20:14, Pierre Pronchery khor...@defora.org wrote:
  There are instructions here:
  https://www.netbsd.org/docs/guide/en/chap-updating.html#building-as-non-root
 
  As described here, I usually place my regular user in the wsrc group
  when installing a system, and then just:
  # mkdir -p /usr/src /usr/pkgsrc
  # chgrp wsrc /usr/src /usr/pkgsrc
  # chmod g+w /usr/src /usr/pkgsrc
 
 What is the wsrc group, what is it supposed to be for and why isn't
 there any documentation about groups in NetBSD? Or is it there, but I
 haven't found it.

Hi

owner and group of my /usr/src and /usr/pkgsrc are that of my 
admin/build user and completely arbitrary. The user does also 
belong to the wheel group and is in /usr/pkg/etc/sudoers.


David

Re: /usr/pkg on its own partition?

[David Lord]

On 18 Aug 2015 at 9:33, Ottavio Caruso wrote:

 Hello,
 
 is it good practice to have /usr/pgk on its own partition? Would it
 make easier when upgrading a whole system?
 
 Or is it  better to have the whole of /usr on its own?
 
 If anybody has got a full desktop installation working, how much does
 /usr/pkg take as a percentage of the whole system?

my build pc:

filesystem Total   Used   %  Mounted on
  MB MB
/dev/wd0a   2234992  46  /
/dev/wd0e  41008   7622  19  /var
/dev/wd0f  41008733   1  /home
/dev/wd0g   1985 10   0  /var/log
/dev/wd0h  41008  19500  50  /home/sysbuild
/dev/wd0i  23557   8642  38  /usr/pkg
/dev/wd0j   2978   1009  35  /usr/local
/dev/wd0k  46533  25727  58  /var/chroot/pkgbuild
/dev/wd0l   2401  2   0  /tmp
/dev/wd0m   2401  2   0  /altroot

Sysbuild has directories for nbsd6-i386, nbsd7-i386.

Pkgbuild has directories for ($release)/usr/ which has
logs, distfiles, packages, pkgsrc/ src, xsrc   

/etc/mk.conf and ($usr)/profiles outside and in chroot
have appropriate entries.  

My desktops have a simpler scheme as they often have
partitions for alternative operating systems. My laptop
(Compaq 2100) has just had NetBSD upgraded from 6.0 to
7.0RC3 and touchpad buttions no longer function 
correctly.


David

Re: installing rsync

[David Lord]

On 24 Jul 2015 at 8:59, William A. Mahaffey III wrote:

 
 
 I just tried to install rsync using pkgin  I got the following:
 
 
 4256EE1 # pkgin install rsync
 calculating dependencies... done.
 
 nothing to upgrade.
 1 packages to be installed: rsync-3.1.1 (294K to download, 655K to install)
 
 proceed ? [Y/n]
 downloading packages...
 rsync-3.1.1 is not available on the repository
 proceed ? [y/N] y
 installing packages...
 pkg_install warnings: 0, errors: 0
 4256EE1 #
 
 
 Nothing in my messages file, /var/db/pkgin/cache had a rsync.tgz dated 
 today when I ran the command, but it was zero sized. Did I just catch 
 the std. repo at a bad time ? Are there other sanctioned repos ? TIA  
 have a good one.
 
Hi

I'm building from 2015Q2 source on i386. Distfiles has 
rsync-3.1.1.tar.gz whilst packages has rsync-3.1.1nb1.tgz.
Maybe you just need to wait and try again in a day or two.


David

Re: ntpd not correcting system time, offset keeps increasing

[David Lord]

On 13 Jan 2015 at 18:43, scar wrote:

 ntpd does not seem to be doing it's job.  i have Ver. 4.2.6p5 installed
 on 6.1.5 and after 51 days of uptime, the system clock is 5 minutes
 behind and the offset just keeps increasing.  ntpdate seems to work fine
 when the system boots, but ntpd just isn't maintaining the time after boot.
 
 here is my ntpd.conf:  http://pastebin.com/EbxzPBw7
 
 i noticed /var/db/ntp.drift doesn't exist which sounds important to keep
 the time maintained
 
Hi


I'm using ntp-4.2.8p1-beta1 NetBSD/6.1_STABLE

Ntpd needs to be in sync for quite some time before the
driftfile is created/populated. 

What ntpd related lines do you have in /etc/rc.conf

I have:
| ntpdate=YES ntpdate_flags=-4 -b -s
| ntpd=YES
| ntpd_flags=-g -N -p /var/run/ntpd.pid

Note:
(1) you should not have same ntpd options in both rc.conf and
ntp.conf otherwise ntpd might not start. 

(2) ntpdate is now deprecated but I find with above that ntpd
syncs within 5 to 15 minutes which nearly always results in 
no reduction in my pool score after reboots or ntpd restarts,
ie scores stay at 20. 


David

 # ntpq -p
  remote   refid  st t when poll reach   delay   offset
 jitter
 ==
 +Maggie.Telcom.A 132.163.4.1022 u6   64  3770.455  297275.
 19.763
 #
 
 i used to have restrict ntp.arizona.edu nomodify notrap after the
 server line at the end of the conf file but it sounds like those
 settings prevent ntpd from working, so i removed it.  but after
 restarting ntpd, the offset reported by ntpq just keeps increasing
 
 

Re: pf and rpi

[David Lord]

On 2 Oct 2014 at 21:20, Christian Koch wrote:

 On Thu, Oct 02, 2014 at 04:06:22PM +0200, Zoran Kolic wrote:
  I put rpi behind the closet. I see this as a proof of the concept.
  Aside my previous idea (pfctl; shutdown -r +10), I got answer:
sleep sometime; pfctl -d
  My biggest puzzle is syntax for pf. On netbsd it is somewhere about
  version 4.2 openbsd. I cannot say for sure if it supports modulate
  state and else.
  Best regards
  
 Zoran
 
 Seriously, why aren't you using NPF? NPF is the packet filter that is actually
 being developed on and for NetBSD. And I'm positive it'll work just fine on
 Raspberry Pi.
 
 http://www.netbsd.org/~rmind/npf/

Hi

I'm not using npf, perhaps because I'm still using ipf as from
1997, and above page doesn't go anywhere on my system with win98
and seamonkey as (as of 2012). My big desktop, nbsd-6, had cd 
drive failure during system update repartitioning but why should
I have to use that anyway. I can also use ipv6 which ipf handles.


David

Re: hp aio : netbsd won't boot ...

[David Lord]

On 25 Sep 2014 at 11:56, Mayuresh Kathe wrote:

 On 2014-09-25 11:22, Manuel Bouyer wrote:
  On Thu, Sep 25, 2014 at 02:59:11AM +0530, Mayuresh Kathe wrote:
  as i'd mentioned before, the newfs command failed to work during
  installation process.
  
  this doesn't help
 
 yeah, figured, that's why i'd sent the following mail too, somehow it 
 didn't make it through.
 
 - another email sent to the list begins -
 
 okay, i retried installation of 6.1.4 on the system after disabling 
 'uefi' boot option in the bios settings.
 
 the installation stopped at the file system creation step, messages as 
 below;
 
 Status: Command failed
 Command: /sbin/newfs -V2 -O 2 -b 32768 -f 4096 /dev/rwd0a
 Hit enter to continue
 
 newfs: /dev/rwd0a: open for read: Device busy
 
 i had faced a similar problem on a previous (dying) machine, thought it 
 to be because of the system's hdd.
 
 now what i have is a brand new system, don't quite know how to proceed.
 
 any help/pointers would be appreciated. :)
 
Hi

it might not help but on both NetBSD-6/i386 and current/amd64,
I've had the same error when I've previously used the utilities
and had various mounts active, both local partitions and nfs and
then went back to complete the install which then failed. After
rebooting and selelecting the installer there was no problem and
installs completed ok.


David

 

Any one here have Radclock working on NetBSD?

[David Lord]

Hi

I've downloaded radclock-0.3.3 from  
http://www.synclab.org/radclock/ but although some of
the files indicate it might be possible to compile on
NetBSD.I'm unable to get it to build on either 
NetBSD-6/i386 or NetBSD-6.99.44/amd64.

Has anyone ported it to NetBSD?


David

Re: Any one here have Radclock working on NetBSD?

[David Lord]

On 8 Jul 2014 at 20:00, Christos Zoulas wrote:

 In article 53bc3902.15128.1702...@netbsd.lordynet.org,
 David Lord net...@lordynet.org wrote:
 
 Hi
 
 I've downloaded radclock-0.3.3 from  
 http://www.synclab.org/radclock/ but although some of
 the files indicate it might be possible to compile on
 NetBSD.I'm unable to get it to build on either 
 NetBSD-6/i386 or NetBSD-6.99.44/amd64.
 
 Has anyone ported it to NetBSD?
 
 Here's enough to get it compiled.
 
 christos
 
 diff --git a/examples/radclock_capture_packet.c 
 b/examples/radclock_capture_packet.c
 index 1c8c04c..c02e854 100644
 --- a/examples/radclock_capture_packet.c
 +++ b/examples/radclock_capture_packet.c
 .

Thanks, great!

I'll not be able to try it out until Jul 10. 

The claimed improvement over stock ntpd is enormous but 
I'm doubtful that the feedforward solution will adapt to 
wide local temperature variations here. 

My pool servers and lan pcs are below 400us rms offset vs 
my local gps/pps source but I'd like to improve on that.
Original plan was to feed pps to each host but that 
requires additional wiring.
 

David

Re: NetBSD Security Advisory 2014-006

[David Lord]

On 16 Jun 2014 at 17:26, Roy Bixler wrote:

 On Mon, Jun 16, 2014 at 11:49:46AM -, David Lord wrote:

 
  I use the default incremental builds which are quite fast
  after the first pass. Only downside for me is that each of
  my /home/sysbuild/nbsd-ver_arch/ directories needs  20G
  disk space. It's probably possible to run multiple 
  ver/arch from a single directory but my build pc with 
  2G ram ground to a halt with all swap+memory used up.
 
 I'm just getting into this myself, both for reasons of a device driver
 problem I was having (see recent timeout on siside0 thread for
 details) and for the SSL security update.  I use the old-fashioned
 method of building from source with CVS as described in the Guide.  It
 took about the amount of space I would expect until I decided to try
 the live-image option, which adds around 10 Gig. to the space
 requirement.  In contrast, building the iso-image didn't take nearly
 as much of a hit.  I haven't used the sysbuild package, but perhaps
 this is what you're seeing?
 

Hi

Probably but disk space hasn't been any problem for me
over past few years.


Twice weekly builds:
20G  nbsd-6_1386/   tools, 6x kernels, release, iso-image,
install-image, live-image

Weekly builds:
18G  nbsd-6_amd64/  tools, release, iso-image, install-image,
live-image

23G  nbsd-cur_i386/ tools, release, iso-image, install-image,
live-image

18G  nbsd-cur_amd64/tools, 1x kernel, release, iso-image,
install-image, live-image


My internet routers, servers and lan pcs were all nbsd-6/i386
but a couple of lan pcs were amd64 with  3G ram and these
are now running nbsd-cur/amd64.


David

Re: NetBSD Security Advisory 2014-006

[David Lord]

On 16 Jun 2014 at 18:53, Ray Phillips wrote:

 Since there aren't going to be binaries containing the latests fixes 
 to ssh on nyftp.netbsd.org for a while, would someone post step by 
 step instructions to get the updated source code, compile it and 
 replace the faulty pieces of NetBSD please?  The advisory just says 
 Update src and rebuild and install. which is a bit too vague for me.
 
 The machines I'm responsible for are running NetBSD/i386 6.1.4.  It 
 seems the latest vulnerabilites are serious enough that machines 
 shouldn't be left running with them, so I'd rather not wait until 
 6.1.5 is released to repair them.
 

Hi

I've been using sysutils/sysbuild + sysutils/sysbuild-user
from pkgsrc. The package is still broken but only requires
${SYSBUILD_BINDIR=/usr/pkg} to point to /usr/pkg/bin.

I use the default incremental builds which are quite fast
after the first pass. Only downside for me is that each of
my /home/sysbuild/nbsd-ver_arch/ directories needs  20G
disk space. It's probably possible to run multiple 
ver/arch from a single directory but my build pc with 
2G ram ground to a halt with all swap+memory used up.

  
David

 
 Ray

Re: optimum building distribution commands

[David Lord]

On 12 Jun 2014 at 18:57, Riccardo Mottola wrote:

 Hi,
 
 to build my NetBSD distribution from current I do:
 
   ./build.sh -O ../obj -T ../tools -U -x distribution
 sudo ./build.sh -O ../obj -T ../tools -U -x install=/
 
 However, I notice that the second command essentially seems to recompile 
 everything, it takes a looong time (as in hours). Am I doing something 
 wrong, or do I miss some option? I'd ideally build as user and then just 
 install with sudo.
 
 Riccardo

Hi

you also need the -u option to specify MKUPDATE=yes

As from about 2013-09 I've been using sysutils/sysbuild
which uses still broken script sysbuild4cron to schedule 
automatic builds. I tried configuring to build all of 
nbsd-6/i386, nbsd-6/amd64, nbsd-cur/i386 and 
nbsd-cur/amd64 but that ground to a halt due to memory
and swap exhaustion so I split to have four separate 
build directories, 20G each (much duplicated wasted disk
space), and that works very well. I've added a script to 
put kernels and sets into a latest directory which is
then used by a system update script.

David

Re: / was overflows

[David Lord]

On 28 Apr 2014 at 23:18, ( Miwa Susumu ) wrote:

 hi.
 
 / (root) was overflows.
 How do I get to find out, consuming capacity?
 
 
 question
 
 I ran the sysbuild build. And it was in error.
 
 / was full for a long time. or impact of sysbuild. I do not know it.
 As it, Where is  / overflowing? I should I examine how?
 
 
 environment
 
 % uname -msr
 NetBSD 6.1.2 i386
 
 
 sysbuild error
 
 % sysbuild build
 
  :
 #   compile  libpuffs/callcontext.ln
 CC=/root/sysbuild/i386/tools/bin/i486--netbsdelf-gcc
 /root/sysbuild/i386/tools/bin/i486--netbsdelf-lint -chapbxzFS -S -w -d
 /root/sysbuild/i386/destdir/usr/include  -D_FORTIFY_SOURCE=2-i
 /usr/src/lib/libpuffs/callcontext.c
 
 /: write failed, file system is full
 


Hi

I've been using sysbuild twice a week as from Oct 5, 2013. You 
might find both memory and disk usage is more than you expected.

Build pc, 3 GB ram, disk usage:
  13G   /home/sysbuild/nbsd-6_i386/i386
 1.1G   /home/sysbuild/nbsd-6_i386/release
 1.5G   /home/sysbuild/nbsd-6_i386/src
 plus logs etc.

Trying to run a conf with 4 x ver + arch ran out of memory and 
swap. Build of a single ver + arch doesn't touch swap.


David

Re: Making a localized, educational live-usb version of Netbsd? Possible, and is worth it?

[David Lord]

On 6 Apr 2014 at 15:27, Eric Haszlakiewicz wrote:

 On April 6, 2014 1:52:29 PM EDT, Aleksej Saushev a...@inbox.ru wrote:
 Eric Haszlakiewicz e...@nimenees.com writes:
 
  On April 6, 2014 7:33:34 AM EDT, Aleksej Saushev a...@inbox.ru
 wrote:
 Ottavio Caruso ottavio2006-usenet2...@yahoo.com writes:
 
  On 5 April 2014 16:56, Aleksej Saushev a...@inbox.ru wrote:
  LiveCD is of no use to people who have no functional CD drive
  or no CD drive at all. This is why it is the wrong approach.
 
  By livecd I meant any system which is not installed to local hard
 disk
  and resets itself after reboot. It doesn't have to boot from a CD,
 it
  can boot from any removable media, the principle is the same.
 
 Live CD is significantly different from live USB pen drive and SD
 card,
 it has to be in a separate category because building it is based on
 completely different principle. Same for live DVD.
 
  How is it any different? In both cases you create a boot image
  that you don't change, and boot it on a machine whose existing

Hi

I must be from a different world?

My usb drives have been writeable, at least from a root
login, and I accidentally destroyed my first one. That is
a big disadvantage but not as important as being able to 
just change config or update as one does with NetBSD on hdd.


David
 

  installation you don't change. That seems pretty similar to me.
  The fact that in one case the media physically prevents you from
 modifying it seems largely irrelevant.
 
 It is relevant and this is exactly what constitutes major difference
 when you work on such a project. You don't need to think how to handle
 various things that require data persistency even if temporal one.
 Your /tmp and /var are writable from the very beginning as if you are
 using HDD. This alone makes USB pen drive a lot different to place it
 into another category.
 
 Well, you clearly have a different idea of what a live usb system is than I 
 do.  I expected it to be something that could be booted repeatedly and have 
 the exact same environment each time (i.e. no changes allowed to the usb 
 storage), while you seem to be talking about just a normal install on a usb 
 stick.
 
 Even if you decide to install some additional software, you don't need
 to create the image from scratch, with USB pen drive you follow the
 usual
 routine.
 
 How the image is created seems to me like something that is not at all 
 relevant to how it is used, and I imagine that 99% of the time will be people 
 *using* a pre-made image, not creating their own.  The times that I've used 
 livecd images the grab-it-and-go feature has been the biggest reason for 
 using it, so I don't care what steps might have been needed to create it.
 
 Eric
 

Re: Making a localized, educational live-usb version of Netbsd? Possible, and is worth it?

[David Lord]

On 4 Apr 2014 at 21:45, Ottavio Caruso wrote:

 On 4 April 2014 21:13, Aleksej Saushev a...@inbox.ru wrote:
  2) Can I start for an existing Netbsd-based live distro? Apparently
  I've been told it's not a good idea.
 
  Sorry? It looks like a misinterpretation, since I didn't see this.
  In fact, I'd say that it would be great to try it out.
 
 Then I misunderstood what you wrote here:
 http://mail-index.netbsd.org/netbsd-users/2014/03/27/msg014342.html
 
 I thought you meant it wasn't a good starting point. But I see you
 have a project going on, so I'm looking forward to give a look at it
 when it's ready.
 

My attempts at an install system from a usb stick
eventually worked well and was useful on my systems
without cdrom drive. Unfortunately I managed to 
overwrite it. There were two downsides to my usb
stick, although on a 16 G stick I only had about 
500 MB of space with install system taking up most
of that, secondly it was painfully slow.

I've just now reinstalled the stick with 5 GB used
for the filesystem:

Filesystem   Size  Used  Avail
root_device  4.9G  443M   4.2G

It needs some cleaning up but it boots ok.

I also have NetBSD-6.1_STABLE-i386-live-sd0root.img.gz
built from sysbuild, that can be transferred using 'dd'.


David

Re: Making a localized, educational live-usb version of Netbsd? Possible, and is worth it?

[David Lord]

On 6 Apr 2014 at 4:12, Chris Bannister wrote:

 On Sat, Apr 05, 2014 at 09:24:31AM -, David Lord wrote:
  
  My attempts at an install system from a usb stick
  eventually worked well and was useful on my systems
  without cdrom drive. Unfortunately I managed to 
  overwrite it. There were two downsides to my usb
  stick, although on a 16 G stick I only had about 
  500 MB of space with install system taking up most
  of that, secondly it was painfully slow.
  
  I've just now reinstalled the stick with 5 GB used
  for the filesystem:
  
  Filesystem   Size  Used  Avail
  root_device  4.9G  443M   4.2G
 
 Can it be extended to use the whole 16G? That could be handy as a
 transportable Live system.

Hi

No problem, I intended adding a couple of extra fdisk
partitions.


  It needs some cleaning up but it boots ok.
  
  I also have NetBSD-6.1_STABLE-i386-live-sd0root.img.gz
  built from sysbuild, that can be transferred using 'dd'.
 
 And this can be dd'd to a USB stick as a bootable live system? Is it
 possible to extend the image to use all the space on the USB stick?

I've not found an option to set the image size but dummy
files could be added to the iso directory and then once
the image is dd'd to the usb stick it should be possible 
to delete the dummy files to create some free space. 

Having played with the usb stick for a while, I'd say 
its use would be limited to doing system updates since
writes to the stick are very very slow.


I've also tried mklivecd which has an iso directory the
contents of which can be changed before generation of the
iso file. I've not yet burnt a cd to test it.


David

 
 -- 
 If you're not careful, the newspapers will have you hating the people
 who are being oppressed, and loving the people who are doing the 
 oppressing. --- Malcolm X

Re: NetBSD as a TimeCapsule?

[David Lord]

On 18 Mar 2014 at 16:50, Fredrik Pettai wrote:

 
  Just a reflection on NetBSD's mdnsd(8):
  
  NetBSD's mdnsd(8) (version 212.1) was imported into NetBSD in September 
  2009.
  My old Snow Leopard (which is probably one of few left) is running version 
  258.21, so the version imported into NetBSD originates from Leopard. So 
  mdnsd in NetBSD seems more than obsolete. If no one plan to upgrade it, 
  maybe it should be removed from the next major release of NetBSD ?
 
 I was informed that a newer version of mdnsd(8) is available in 
 pkgsrc/net/mDNSresponder, so I installed that to see if that would help, but 
 it fails to start with the message:
 
 Mar 18 14:56:53 xsrv1 mDNSResponder: mDNSResponder (Engineering Build) (Mar 
 18 2014 14:43:05) starting
 Mar 18 14:56:53 xsrv1 mDNSResponder: ERROR: bind(listenfd, (struct sockaddr 
 *) laddr, sizeof(laddr)); failed: 48 (Address already in use)
 Mar 18 14:56:53 xsrv1 mDNSResponder: ERROR: udsserver_init: 48 (Address 
 already in use)
 Mar 18 14:56:53 xsrv1 mDNSResponder: mDNS_AddDNSServer: Lock not held! 
 mDNS_busy (0) mDNS_reentrancy (0)
 Mar 18 14:56:53 xsrv1 mDNSResponder: mDNS_AddDNSServer: Lock not held! 
 mDNS_busy (0) mDNS_reentrancy (0)
 Mar 18 14:56:53 xsrv1 mDNSResponder: mDNSResponder (Engineering Build) (Mar 
 18 2014 14:43:05) stopping
 
 and the bundled dns-sd doesn't seem to work well either:
 
 xsrv1# /usr/pkg/bin/dns-sd -V
 DNSServiceGetProperty failed -65563
 
 (still testing on the NetBSD 6.1.3/amd64 host...)
 
 So to iterate over this again, is anybody using mDNSresponder successfully on 
 NetBSD? (either the bundled in NetBSD, or the pkgsrc version...)
 
 /P
 

Hi

My amd systems aren't usable just now and have been running 
i386 anyway. On an i386 system mdnsd fails to start but after
adding user _mdnsd and group _mdnsd: 

$ /etc/rc.d/mdnsd onestart
$ ps ax | grep mdnsd
6020 ?   Ss  0:00.01 /usr/sbin/mdnsd


I've no idea how to test if it does anything useful


David

Re: Muddled wrt pkgsrc

[David Lord]

On 10 Mar 2014 at 15:43, Bob Bernstein wrote:

 After a long delay I have finally upgraded my netbsd machine 
 to 6.1.3, and refreshed the /usr/pkgsrc tree to 
 pkgsrc-2013Q4. I then charged off to build claws-mail and 
 ran into a bit of pesky business that has always baffled me.
 
 At every turn, as claws-mail ground through its build, I was 
 confronted with the need to pkg-delete older versions of its 
 many, many dependencies.
 
 Is there a way to tell pkgsrc simply to 'pkg_delete -f' all 
 these older versions of dependencies as and when the build 
 wishes to replace them with uptodate versions?
 

I run pkg_chk -uan to get an idea what packages have
moved and then for those pkg_delete -r pkglist or
pkg_delete -rR pkglist.

I've a script 'update-pkgsrc.sh' from Greg Troxel that
might be useful. 

I mostly build in a diy chroot. My method works well 
enough with almost 2000 packages. I then install to 
target pcs from the built packages using 'pkg_chk -buak'
which can have a few failures sometimes requirng build 
from source on the target. Some of the failures are due 
to my pcs having some uid/gid conflicts.


David

 
 Muddled in Rhode Island
 
 
 -- 
 IMPORTANT: This email is intended for the use of the individual
 addressee(s) named above and may contain information that is
 confidential, privileged or unsuitable for overly sensitive
 persons with low self-esteem, no sense of humour or irrational
 metaphysical beliefs.
 

Re: Broken system after pkgin full-upgrade

[David Lord]

On 20 Jan 2014 at 9:51, Ottavio Caruso wrote:

 On 20 January 2014 00:54, Bob Nestor rnes...@tx.rr.com wrote:
  One thing I did see is if the upgrade involves a lot of packages pkgin
   will calculate the amount of space it needs in /var and reject the
   operation if there isn't enough space
 
 I don't recall receiving errors regarding space. All the packages were
 downloaded at once. Only half way through installation it complained
 that some packages could not be installed. Unfortunately I cannot find
 a log of these errors.
 
 It this can help, I recall viewing these messages:
 File  can be safely deleted if no other applications are using it.
 
 I am just recalling from memory.
 
 Isn't pkgin developed by the core developers at Netbsd. If there are
 any risks of running a full upgrade maybe that option should be
 removed. I can't see what I've done wrong before launching the
 command.
 
 Is there any way to reset the /usr/pkg and start again? At this point
 I am not even sure the output of pkg_info can be trusted.
 

Does output from pkg_info -a look ok?

It's possibly worth doing 
pkg_admin check

and if that throws up errors

pkg_admin reuild-tree or even pkg_admin rebuild

I don't use pkgin, I edit a pkgchk.conf then use pkg_chk
to build/install binary packages (takes 1-3 weeks). The
binaries in /usr/pkg_tarup are then copied to target pcs
and installed using pkg_chk -buak (takes a few hours).


David

 
 
 -- 
 Ottavio

Re: Cable TV on the PC with Netbsd

[David Lord]

On 18 Dec 2013 at 13:40, Ottavio Caruso wrote:

 On 18 December 2013 12:48, Benny Siegert bsieg...@gmail.com wrote:
  If you buy a digital cable set-top box, it usually comes with a way of
  getting the TV stations streamed as MPEG2 streams. Then, you can use
  VLC to watch TV.
 
 What about analogue TV instead? Which drivers do I need?

Hi

I'm still using a TV card on Win98 but other PCs have 
similar TV cards with ancient installs of SuSE, FreeBSD
and NetBSD. What works is very dependent on the whole
system, cpu, motherboard chipset, operating system and
software versions. On one setup I was reliably running
FlightGear from NetBSD's linux emulation but neither
FreeBSD nor native SuSE on same pc was reliable. TV was
also working from NetBSD on that system.
 
You need to know the chipsets to answer what drivers you
need then in most cases NetBSD will already have detected 
and be able to use the TV card. In some cases the card 
will be detected but not configured due to a variation in
chipset version which can be from trivial to impossible 
to fix.
  

David

 
 -- 
 Ottavio

Re: Frustrating network latency issue on NetBSD 6.1

[David Lord]

On 20 Jun 2013 at 20:36, Mayuresh wrote:

 On Thu, Jun 20, 2013 at 12:45:08PM -, David Lord wrote:
   I think 36% is a large number for packet loss?
  
  It is unless the link is very busy when icmp might be 
  dropped.
 
 Ok, I'll check whether there is a way in this router to check its network
 activity.
 
  What was your ping commandline?
 
 Just ping router where router is router's host name.
 
 Mayuresh.

If the problems only started since you upgraded to NetBSD-6.1 
that suggests either a misconfiguration or an incompatibility
has been introduced.

Have you compared /var/boot/dmesg.boot and /var/log/messages
for errors from before and after you upgraded?


David

Re: Starting service (particularly ntpd) in background

[David Lord]

On 26 May 2013 at 15:01, Mayuresh wrote:

 On Sun, May 26, 2013 at 12:46:19AM -, David Lord wrote:
  My own /var/log/ntp/ntp.log has nothing and /var/log/messages
  just confirms I don't have a problem so maybe the OP should
  give some more details.
 
 I realize, the time taken is not by ntpd, but by ntpdate. I have both
 enabled in my rc.conf.
 
 But ntpdate does take something close to 20s.
 
 The reason I keep ntpdate enabled is, the system is also used with an
 alternative OS (Linux) and sometimes the clock is found offset by GMT to
 my time zone's difference. Starting just ntpd just gives up and does not
 correct this difference.

I think that is a problem that you should fix. I also
multiboot with some systems BSD, Linux and Windows.
I have them all set to UTC. BSD and Linux can apply 
timezone correction and display local time but ntpd 
in both cases is UTC. 

 
 Is it alright to keep ntpdate enabled by default? Why does it take so
 long?

I have ntpdate_flags=-4 -b -s as otherwise ipv6 might 
be tried and fail (only one of my servers has ipv6
access to internet just now).
 
I sometimes have to reboot my servers a couple of times 
for NetBSD updates and use of ntpdate significantly 
reduces time to get my ntp service synced and usable.

If you don't need better than about 50 msec or so accuracy 
immediately after bootup you can set ntpdate=NO in rc.conf.


David

 
 Mayuresh.

Re: Starting service (particularly ntpd) in background

[David Lord]

On 25 May 2013 at 22:45, Mayuresh wrote:

 I find ntpd to be time taking during the boot process. There will of
 course be some explanation of this. I can also possibly try using nearby
 servers/switches that support ntp to reduce the time.
 
 However ntpd is a service that I afford to start even after the boot. It
 does not need to delay my getting a login prompt.
 
 Is there any way to start a service in the background in NetBSD?
 Alternatively, is there any way to configure ntpd to start in background?
 
 Mayuresh

Hi

I've never noticed a significant delay.

I have ntpdate run first before ntpd as that gives a 
quicker time for ntpd to be in sync with the minimum 
of 3 sources from those in ntp.conf.

From /var/log/messages: 
ntpdate takes  6 sec.
ntpd takes  2 sec

This is from a server running NetBSD-6.1-stable i386.


David

Re: Starting service (particularly ntpd) in background

[David Lord]

On 25 May 2013 at 22:51, Rhialto wrote:

 I think the way ntpd is started the way it is, is to get proper
 timestamps in logfiles, and no time jumps when services are running.
 So it is likely put as early in the boot process as possible.
 
I've no problem with that and ntpd here starts faster than 
many other services I run that are started via rc.conf.

My own /var/log/ntp/ntp.log has nothing and /var/log/messages
just confirms I don't have a problem so maybe the OP should
give some more details.

The answer to his question is probably to disable ntpd in
rc.conf and run a script from rc.local but that would just
move the problem elsewhere.


David

Re: Status of NetBSD 6.1_RC3?

[David Lord]

On 28 Apr 2013 at 5:32, Thomas Mueller wrote:

 What is the status of NetBSD 6.1 RC3?  I see references to RC3 in some 
 messages, but not on
 ftp://ftp.netbsd.org/pub/NetBSD/ .
 
 I noticed on nyftp.netbsd.org that there was a build of 6.1_RC3, but don't 
 know what was the final verdict.
 
 I update the source by cvs and notice reference to RC3 in a CHANGES file, but 
 being in the source tree does not mean accepted for release or release 
 candidacy.
 
 Tom
 

I last updated netbsd-6 on April 17, 2013.

$ uname -v
NetBSD 6.1_RC3


David

Re: Is it possible to chroot and then install NetBSD to the boot drive

[David Lord]

On 19 Mar 2013 at 8:51, Al Zick wrote:

 Hi,
 
 On Mar 17, 2013, at 6:57 PM, Greg Troxel wrote:
 
 
  Al a...@familysafeinternet.com writes:
 
  What I would like to be able to do is do an install on either another
  drive or on an NFS partition. Then chroot to it and then delete the
  files that are on /, /var, and /usr and then do a clean install of
  NetBSD. Is this possible? If it is, could someone give me a how to on
  doing something like this?
 
  chrooting will be awkward, because once you chroot then you will be
  running the binaries from the new install, which only works if your  
  host
  system is compatible (NetBSD, same or later, more or less).  And if  
  you
  remove files, you won't be able to run new commands.
 
  However, you don't need to chroot.  If you take a new drive, and
  fdisk/disklabel it, and mount it, and then unpack the sets, and then
  install bootblocks, it should work.  You didn't mention, but if you  
  mean
  i386, then bootblocks consists of installboot for bootxx_ffsv1
  (probably) and also /boot, plus mbr boot records.  If the host is  
  NetBSD
  you can just use installboot and cp, pointing to /usr/mdec in the new
  system.
 
  I have done something similar a number of times, basically moving a
  system from old disks to new (bigger, less aged and in theory more
  reliable) disks.  So instead of unpacking sets, I have done dump/ 
  restore
  or rsync from old to new (under /mnt) and then installed bootblocks.
 
 I have a system that I sent new drives out to the data center not  
 that long ago. I did a new install of NetBSD on them, but then I  
 decided to update to NetBSD 6.0. In the process, I broke some things.  
 Now the system is not very usable. I can get another hard drive and  
 install NetBSD on it and then send them that drive, but someone at  
 the data center will need to swap out the drive again. I don't see  
 how I could netboot this system and it is not real easy to tell it to  
 boot from another drive or partition, so I thought that there might  
 be some way to setup a root filesystem much like you do when you  
 netboot. Temporarily mount this partition to run make dev files, then  
 switch over to it, and unmount /home, /var, /usr,  and finely /. Run  
 newfs on the old /, /var, and /usr. Mount / as /mnt, then /var  as / 
 mnt/var and /usr as /mnt/usr. Then cd to /mnt and untar a working  
 install that I did on a test box that has the exact same config as  
 this system, so that everything will be right in /etc. Is there any  
 way to do this?
 
 Thanks,
 Al
 

Assuming you still have remote access to the system  it may
not require the above and you can try method below on your
test setup.

I had an unusable system when I had tried update from 5.x 
to 6.0-beta and then tried to backout to 5.y. Possibly I
had a bad 6.0beta cdrom. I then found that backing out is
not supported.
 
I still had ssh access and eventually I copied a new kernel
and tried to use a script to install sets. This is best done
single user but would require network access to be enabled
and I've not tried that as my pcs are all accessible. 
Stopping as many services as practical may be advisable.

(1) copy sets to some directory

(2) copy new kernel to /

(3) copy script to some directory 

(4) reboot

(5) unpack sets
#!/bin/sh
cd /
for i in base comp games man misc modules tests text ; do
pax -rvzpoe -f /sets/${i}.tgz
done
#

(6) reboot

The etc.tgz set is deliberately missing from above 
If feeling brave you could then use /etcmanage to tidy up
but I prefer to use an occasional sysinstall from a cd.

Without using /etcmanage I've updated from 4.x through to
6.1RC1 but have other systems with more recent cd installs
and along with warnings in /var/log/messages or 
/var/run/dmesg can fix problems in /etc later.

With 4.x and possibly 5.0 I needed to run fsck_ffs -c 4
for all the filesystems and change entries in fstab to
enable wapbl. 


David